yfedoseev/pdf_oxide

The fastest PDF library for Python and Rust. Text extraction, image extraction, markdown conversion, PDF creation & editing. 0.8ms mean, 5× faster than industry leaders, 100% pass rate on 3,830 PDFs. MIT/Apache-2.0.

What it solves

PDFOxide is a high-performance PDF toolkit designed to overcome the speed and licensing restrictions of existing libraries. It provides a fast, MIT-licensed alternative to tools like PyMuPDF, enabling efficient text, image, and layout extraction from PDFs across 20 different programming languages.

How it works

The project is built with a Rust core that handles the heavy lifting of PDF processing. To provide broad accessibility, it exposes a stable C ABI, which allows it to create idiomatic bindings for 19 other languages (including Python, Go, JS/TS, Java, and C++). It also includes a CLI tool for terminal-based processing and an MCP server that allows AI assistants like Claude and Cursor to interact with PDFs locally.

Who it’s for

  • RAG and LLM developers who need to convert PDFs to clean Markdown for retrieval-augmented generation pipelines.
  • AI assistant builders who want to integrate local PDF access into their agents.
  • Data engineers processing thousands of PDFs at scale for academic research or structured data extraction from forms and tables.
  • Software developers seeking a PyMuPDF alternative with a permissive MIT license.

Highlights

  • Extreme Speed: 0.8ms mean processing time per document, claiming to be 5x faster than PyMuPDF and 15x faster than pypdf.
  • Broad Language Support: Native bindings for 20 languages, including Rust, Python, Go, JS/TS, C#, Java, Kotlin, Swift, C++ and more.
  • High Reliability: 100% pass rate on a corpus of 3,830 real-world PDFs (veraPDF, Mozilla pdf.js, and DARPA SafeDocs).
  • AI-Ready: Includes a built-in MCP server for direct integration with AI assistants.
  • Comprehensive Feature Set: Supports text, image, and layout extraction, as well as PDF creation, editing, and form filling.